Everyone Laughs Before They Lose
From the 48 Laws of Power PAGE-MAX re-ingest, Laws 47–48. Three instances, all in source, none from recall.
Cyrus, at the river Araxes. Queen Tomyris has written to him twice. The second letter offers him terms with his victory intact: give back my son, leave with your forces whole, "be content with your triumph over a third part of the Massagetai." Greene: "Cyrus scoffed at her."[^1] He is dead within the chapter, head in a wineskin.
Xerxes, before Greece. His adviser Artabanus tells him the two mightiest powers in the world are against him. "Xerxes laughed—what powers could match his gigantic army?" Artabanus names them: the land and the sea. No harbour big enough for the fleet, no supply line long enough for the army. Greene: "Thinking his adviser a coward, Xerxes proceeded with the invasion."[^2] Both predictions land exactly.
Chiang's staff, in Manchuria. The Communist units are dispersing in ways nobody can read. "Commanders comfortably remote from the front lines might laugh at Mao, but the soldiers had fought the Communists in the mountains, and had come to fear their elusiveness."[^3] Within three years the Nationalists have lost China.
Three warnings. Specific, correct, delivered before the decision, all available in time.
Greene supplies all three and never puts them together, and the standard reading is the one you'd expect from him — arrogance, hubris, the intoxication of victory. That reading has a problem: it names a state, and states aren't observable from inside. Nobody has ever caught themselves being arrogant in the moment. It doesn't feel like arrogance. It feels like clarity.
So look at what actually happens instead, because it isn't a feeling. It's a move, and it's the same move all three times.
Artabanus makes a claim about harbour capacity and supply lines. Those are numbers. They can be checked. Xerxes doesn't check them — he decides Artabanus is a coward.
Watch what that does. The claim was about the world. Now it's about the man. And a claim about the man cannot be answered with facts about harbours, because facts about harbours are no longer what's being disputed. The disagreement has been moved somewhere the evidence can't reach.
Tomyris gets the same treatment — she becomes a barbarian queen making noise, rather than the person best placed on earth to price a third of her own army. Mao becomes more a poet and philosopher than a general.[^4]
⚑ The fatal act isn't feeling superior. It's reclassifying a factual claim as a character claim — and unlike the feeling, that one has a timestamp, a sentence attached, and usually a witness.
Every framework about failing at the top of your game is aimed at a state, and the thing that's actually observable is a speech act.
Hubris, complacency, the winner's curse, overconfidence bias — all internal, all requiring you to detect your own condition using the apparatus the condition has compromised. That's why the advice never lands. Be humble after a win is delivered to a person for whom humility, right now, feels like a failure of nerve.
The reclassification is different in kind. It's outside you. It happens in a sentence, in a room, at a time. She's being emotional. He's not ambitious enough. They don't understand what we're doing. He's a coward. Someone brings you a number and you answer with an adjective about the person.
⚑ And it's cheap to catch, because you don't need to assess yourself at all — you only need to notice a grammar. Did my response engage the claim, or did it describe the claimant?
There's a second half, and Manchuria has it. The correct assessment usually exists and is held at the wrong altitude. The Nationalist soldiers were frightened; the commanders laughed. Nobody was missing information. The information was in the building, and the gradient ran the wrong way — contempt highest where contact was lowest.
Which turns a mood into two things you can measure. Ask whether your organisation's read of an opponent varies by seniority. Ask whether your last response to a warning was about the warning. Neither requires you to have an accurate view of yourself, which is fortunate, because at the moment this matters you won't.
- Three cases picked by an author for their outcomes is not a base rate, and this is the finding's real weakness. Greene selected Cyrus, Xerxes and Chiang because they lost. Rulers dismiss cautious advisers constantly and most of the time nothing happens — sometimes because the adviser was, in fact, timid. Nothing here establishes how often the same move correctly discards bad advice, and until that's known the pattern is a shape, not a rate.
- The falsifier is specific: find documented cases where a leader recategorised a warner and was right to. If those are common, the move is a normal and necessary filter and its presence in three disasters is survivorship.
- The altitude gradient is separately checkable and much easier. It doesn't need history — it needs one organisation, two levels, and the same question about a competitor. If the front line is consistently more worried and consistently more accurate, that's a live finding you can act on this quarter.
- Named candidate sources: Janis on groupthink, which has the mechanism at the group level and a documented case set · Tetlock's forecasting work, for whether dismissed warnings outperform accepted ones · Herodotus directly for Artabanus, since Greene cites nobody · and a modern intelligence-failure case study — Pearl Harbor or the Yom Kippur War — where the warning-that-didn't-travel-upward is the entire literature.